International Conference on Computer and Knowledge Engineering

Home / 14th International Conference on Computer and Knowledge Engineering

Non-Negative Matrix Factorization improves Residual Neural Networks

Authors :

Hojjat Moayed¹

1- Esfarayen University of Technology

Keywords :

ResNet،Residual Neural Network،NMF،Deep Learning

Abstract :

Residual neural networks enable the use of very deep architectures. These architectures benefit by passing the identity information from a layer directly to subsequent layers. Extensive research has been conducted to improve the performance of residual neural networks. In this paper, we propose a method to improve performance by providing a more informative input using non-negative matrix factorization. The method combines the invariant features learned from the training data with the extracted, fine-tuned features at the end of the residual block. Our experimental results confirm that the proposed architecture improves performance on the image classification task.

List of archived papers

Spatial-channel attention-based stochastic neighboring embedding pooling and long short term memory for lung nodules classification

AHMED SAIHOOD - HOSSEIN KARSHENAS - AHMADREZA NAGHSH NILCHI

Spatio-Temporal Graph Neural Networks for Accurate Crime Prediction

Rojan Roshankar - Mohammad Reza Keyvanpour

ExaAEC: A New Multi-label Emotion Classification Corpus in Arabic Tweets

Saeed Sarbazi-Azad - Ahmad Akbari - Mohsen Khazeni

Delay Optimization of a Federated Learning-based UAV-aided IoT network

Hossein Mohammadi Firouzjaei - Javad Zeraatkar Moghaddam - Mehrdad Ardebilipour

Distilled BERT Model In Natural Language Processing

Yazdan Zandiye Vakili - Avisa Fallah - Hedieh Sajedi

Pruning and Mixed Precision Techniques for Accelerating Neural Network

Mahsa Zahedi - Mohammad Sediq Abazari Bozhgani - Abdorreza Savadi

Histopathology Image-Based Cancer Classification Utilizing Transfer Learning Approach

Amir Meydani - Alireza Meidani - Ali Ramezani - Maryam Shabani - Mohammad Mehdi Kazeminasab - Shahriar Shahablavasani

Efficient Object Detection using Deep Reinforcement Learning and Capsule Networks

Sobhan Siamak - Eghbal Mansoori

Towards Efficient Capsule Networks through Approximate Squash Function and Layer-wise Quantization

Mohsen Raji - Kimia Soroush - Amir Ghazizadeh

MIPS-Core Application Specific Instruction-Set Processor for IDEA Cryptography − Comparison between Single-Cycle and Multi-Cycle Architectures

Ahmad Ahmadi - Reza Faghih Mirzaee

more

Samin Hamayesh - Version 42.2.1